Enhanced accumulation of catharanthine and vindoline in Catharanthus roseus hairy roots by overexpression of transcriptional factor ORCA 2

Plant Biotechnology Research Center, Key Laboratory of Urban Agriculture (South), Ministry of Agriculture, FudanSJTU-Nottingham Plant Biotechnology R & D Center, School of Agriculture and Biology, Shanghai Jiao Tong University, Shanghai 200240, China. State Key Laboratory of Genetic Engineering, School of Life Sciences, Fudan-SJTU-Nottingham Plant Biotechnology R & D Center, Morgan-Tan-International Center for Life Sciences, Fudan University, Shanghai 200433, China. Department of Biochemistry and Molecular Biology, Shenyang Medical College, Shenyang 110034, China.

reduce hypertension (van der Heijden et al., 2004).Most of these TIAs are produced at low levels in the natural plants and are difficult to be chemically synthesized due to their complex structures.The need for chemotherapy treatment of cancers has prompted extensive efforts to develop inexpensive and efficient approaches for production of these TIAs.
TIAs biosynthetic pathway in C. roseus is complex with multiple steps and is under strict molecular regulation (Liu et al., 2007).A variety of TIAs are derived from stricto-sidine which is condensed from tryptamine and secologanin and catalyzed by strictosidine synthase (STR).The tryptamine is converted from tryptophan by tryptophan decarboxylase (TDC) through the Shikimate pathway (indole pathway), while the secologanin is derived from geraniol via the terpenoid pathway.Strictosidine can be modified by strictosidine glucosidase (SGD) to form cathenamine, a precursor to various biologically active alkaloids (Collu et al., 2001).It was approved that there is an equilibrium between cathenamine and 4,21-dehydrogeissoschizine (Heinstein et al., 1979).Facchini and St-Pierre (2005) pointed out that many important monoterpenoid indole alkaloids, such as tabersonine and catharanthine, are produced via 4,21-dehydrogeissoschizine, but their enzymology has not been established.The biosynthetic route from 4,21-dehydrogeissoschizine to tabosonine is not completely confirmed.However, it has been established that tabersonine is transformed to vindoline through a sequence of six enzymatic steps.In the final dimerization step of the TIAs biosynthetic pathway, a class III basic peroxidase (CrPRX1) was approved to catalyze the coupling of the monomeric precursors vindoline and catharanthine into -3',4'-anhydrovinblastine (AVLB), the common precursor of all dimeric alkaloids.AVLB is then transferred into VLB and VCR (Sottomayor et al., 1998).
Promoter analysis of the genes that encode STR and TDC reveals that both contain sequences involved in the regulation by stress signals such as UV-irradiation and fungal elicitors (Ouwerkerk et al., 1999;Pasquali et al., 1999).Two transcription factors were isolated by yeast one-hybrid screening with Str promoter sequence (Menke et al, 1999).The two proteins were called ORCA1 and ORCA2, for octadecanoid-responsive Catharanthus AP2/ ERF-domain protein (ORCA).Co-transformation experiments showed that transient overexpression of ORCA2 activated the Str promoter, whereas overexpression of ORCA1 had little effect on Str promoter activity.Transient expression assays also indicated that ORCA2 transactivated the Str promoter via direct binding and its expression was rapidly inducible with jasmonate (JA) and elicitor, whereas ORCA1 was expressed constitutively.Considering that STR is a very important enzyme in TIAs biosynthesis and that STR activity is controlled by expression of ORCA2, therefore, it is necessary to verify the relationship the between function of ORCA2 and TIAs biosynthesis in C. roseus cells.
In this study, transcription factor Orca2 gene was transformed into C. roseus hairy root cultures to investigate the transgenic effect of overexpression of ORCA2 on the TIAs biosynthesis in C. roseus hairs roots.The results showed that the transgenic hairy root extracts accumulated more catharanthine and vindoline in comparison with the control hair root lines, but VLB could not be detected in the transgenic and non-transgenic hairy root cultures by HPLC analysis.The reasons for the results are discussed.Liu et al. 3261

Construction of plant expression vector
The -glucuronidase (GUS) expression cassette was excised from the plasmid pBI121 by EcoRI and HindIII (New England Biolabs, USA) double digestion, and then integrated into plasmid pCAMBIA 1304 (CAMBIA, Canberra, Australia) to form pCAMBIA1304 + (p1304 + ).The plasmid p1304 + contains two GUS expression cassettes and a hygromycin-resistant gene (hpt) expression cassette, which are all driven by CaMV35S promoter.
Total RNA was isolated from the one month old seedlings of C. roseus by CTAB method (Chang et al., 1993).Using oligo (dT)18 as template primer and the total RNA as template, the first strand of cDNA was synthesized by AMV reverse transcriptase (Takara Company, China), and then the second strand of the cDNA was replicated by Escherichia coli DNA polymerase I after cutting mRNA into oligonucleotides with RNaseH (Takara Company, China).
Based on the CDS sequence of Orca2 gene (Genbank Accession No. AJ238740), forward primer FO2 5'-GAAGATCTATGTATCAA TCAAATGCCCATAATTCC-3'(Bgl II) and reverse primer RO2 5'-GGGTCACCTTATTGAGGACGAAGATGACACG-3'(BstE II) were designed and synthesized for the amplification of Orca2 gene.The primer sequences contained BglII or BstEII restriction endonuclease site separately.C. roseus cDNAs were used as template and the PCR was performed with PrimeSTAR ® HS DNA polymerase kit (Takara Company, China) by denaturing at 94°C for 1 min, followed by 30 cycles of amplification (98°C for 10 s, 55°C for 15 s and 72°C for 60 s) and then a 10 min final extension at 72°C.The PCR fragment was then cloned into plasmid pMD18-T (Takara Company, China) to form plasmid pMD18-T-Orca2.After confirming the sequence, the Orca2 gene was excised from pMD18-T-Orca2 by BglII / BstEII double digestion.
Lastly, plant expression vector p1304 + -Orca2 was constructed by replacing a gus gene in the plasmid p1304 + with the Orca2 gene by BglII / BstEII double digestion.The vector p1304 + -Orca2 was transformed into the disarmed Agrobacterium rhizogenes C58C1 strain carrying the plasmid pRiA4 of A. rhizogenes, and the resulting A. rhizogenes C58C1 strains was used for the transformation study.

Cultures conditions and genetic transformation
Seeds of C. roseus, purchased from PanAmerican Seed Company (Cherry Red, USA), were surface sterilized and placed on MS solid medium (1962) in the greenhouse at 25°C for germination.Young leaves from one month old germinated seedlings of C. roseus were broken with sterile surgical knife and pre-incubated on half-strength MS solid medium for 2 to 4 h, and then cultivated with the A. rhizogenes C58C1 strain (OD600 = 0.7) containing vector p1304 + -Orca2 for co-cultivation.After 48 h co-cultivation, the leaves were maintained on the regulator-free half-strength MS solid medium containing 500 mg/l cefotaxime to eliminate bacterial contamination.Two weeks after the C58C1 strain infection, hairy roots were induced from the wounded edges and surface of the leaf explants.Single transformed roots were excised when they grew over 2 cm in length and were maintained separately as independent clone.The hairy root lines were grown at 25°C in the dark and were routinely subcultured to fresh regulator-free half-strength MS solid medium every two weeks.After two months of subculture on solid medium, hairy root cultures were obtained and transferred individually into the regulator-free half-strength B5 liquid medium for continuous subculture.All hairy root cultures were kept at 25°C on a rotary shaker at 100 rpm in the dark.After 30 days of subculture in liquid medium, the hairy root cultures were filtered, washed with 10 ml sterile distilled water and lyophilized immediately in liquid nitrogen for molecular analysis and TIAs extraction.The control hairy root lines were generated by transforming the leaf explants with C58C1 strain lacking vector p1304 + -Orca2 and were grown for subculture as earlier described.

Polymerase chain reaction (PCR) analysis
The bacterium-free hairy root lines were collected, dried on sterile filter paper and quickly frozen in liquid nitrogen.Total genomic DNA was isolated from putative transgenic and control hairy root lines by using the CTAB DNA isolation method (Woodhead et al., 1998).
The DNA samples were then used in PCR analysis for detecting the presence of Agrobacterium rol (rolB, rolC) genes and hpt gene in transgenic hairy root cultures.Oligonucleotide primers for the PCR detection of rolB, rolC and hpt gene were designed based on the DNA sequences of these genes described by Fumer et al. (1986).For the amplification of rol genes (rolB, rolC), primers FrolB (5'-GCTCTTGCAGTGCTAGATTT-3'), RrolB (5'-GAAGGTGCAAGCTA CCTCTC-3'), FrolC (5'-CTCCTGACATCAAACTCGTC-3') and RrolC ( 5'-TGCTTCGAGTTATGGGTACA-3' ) were used.For the detection of the hpt gene, Fhpt (5'-CGATTTGTGTACGCCCGACA GTC-3') and Rhpt (5'-CGA TGTAGGAGGGCGTGGATATG-3') were used.PCR for the detection of all the above genes was carried out by denaturing the template at 94°C for 3 min followed by 35 cycles of amplification (94°C for 40 s, 55°C for 30 s and 72°C for 1 min) and then a 10 min final extension at 72°C.The PCR products were separated by electrophoresis in 0.8% (w /v) agarose gel.Plasmid DNA from A. rhizoenes strain containing plasmid p1304 + -Orca2 was used as positive control and genomic DNA from untransformed C. roseus root was used as a negative control in PCR analysis.

Real-time quantitative analysis (RT-QPCR)
After 30 days of subculture in liquid medium, total RNA was extracted separately from the putative transgenic and control hairy root cultures with plant RNA mini kit (Watson Company, Shanghai, China), and treated with RNase-free DNase (Takara Company, China) to eliminate the potential contaminating residual DNA.The quality and concentration of RNA samples were tested by agarose gel electrophoresis and spectrophotometer analysis.Total RNA was reversely transcribed by using AMV reserve transcriptase (Takara Company, China) to generate cDNA, which was then subjected to RT-QPCR analysis for the expression of Orca2 gene.RT-QPCR was performed on a RoterGene 3000 instrument (Corbett Research, Sydney, Australia).The gene-specific primers FOrca2 (5'-GATCAGGATAATTACGAAGACGAAGT-3') and ROrca2 (5'-AGTTCCCAACCATATCCTCGATCCTT-3') were designed according to the conserved sequence of C. roseus Orca2 gene and were used to amplify the Orca2 gene.Ubiqitin gene (house-keeping gene) was used as an internal calibrator.FUbi (5'-GTGACAA TGGAACTGGAATGG-3') and RUbi (5'-AGACGGAGGATAGCGTG AGG-3') were used as primers to amplify the Ubiqitin gene.The RT-QPCR was carried out with SYBR ® PrimeScript ® RT-PCR kit (Perfect Real Time) according to manufacturer's instructions (Takara Company, China) as follows: 1 min predenaturation at 95°C, 1 cycle; 10 s denaturation at 95°C, 30 s annealing at 56°C and 15 s collection fluorescence at 72°C and 42 cycles.The products of RT-QPCR were run on 1.0% (w/v) agarose gel electrophoresis and it showed an equal-sized band as predicted.Quantification of the gene expression was done with comparative computed tomography (CT) method.The quantitative analysis was repeated three times for each sample.

GUS histochemical assay
GUS assays of the hairy roots were performed by histochemical staining as described by Jefferson et al. (1987) with slight modifications.Both the putative transgenic and the negative control hairy root segments (about 10 mm in length) were incubated at 37°C in the dark in an X-Gluc solution.The X-Gluc solution contained 50 mM Na3PO4 (pH 7.0), 10 mM Na2EDTA, 0.1% (v/v) Triton X-100, 0.1 M K3[Fe(CN)6], 0.1 M K4[Fe(CN)6], 0.25 mM L-1 5bromo-4-chloro-3-indolyl-b-D-glucuronide (X-Gluc) and 20% methanol.The roots were subsequently washed in an ethanol gradient at room temperature (30 min in 70% ethanol, 30 min in 40% ethanol and then 30 min in 20% ethanol).After rehydration, the roots were kept in water and then mounted on a slide for observation and photography.

Alkaloids extraction and high performance liquid chromatography (HPLC) analysis
The 30 day old transgenic and control hairy root samples cultured in half-strength B5 liquid medium were harvested and lyophilized overnight, respectively.The resulting dried roots were weighted, ground to very fine powder using a mortar and pestle, and extracted three times at room temperature with 10 ml of MeOH for 1 h in a sonicating bath.The mixture was centrifuged at 13000 g for 15 min at 15°C (Singh et al., 2000).The supernatant was removed and the biomass was re-extracted again prior to HPLC analysis.
The alkaloid analysis of C. roseus hairy root samples was performed on a Waters Alliance HPLC system (Alliance model 2690; Waters Corporation, Milford, MA, USA) and separated using a C18 column with binary gradient mobile phase profile (55% 5 mmol/l pH6.0 sodium phosphate buffer, 45% acetonitrile) (Singh et al., 2000;Tikhomiroff and Jolicoeur, 2002).Extracts were analyzed by HPLC with a photodiode array detector (Model 996, Waters) to verify the identity and purity of peaks of interest.HPLC with UV detection at a single wavelength only was employed for quantifycation of TIAs.An aliquot of 10 l injection volume provided adequate signal at 220 nm.Authentic standards of catharanthine, vindoline and vinblastine (Sigma, USA) were prepared separately in methanol at a final concentration of 5 g/l and used for the preparation of the calibration graphs.Quantification was repeated three times for each sample.

Establishment and subculture of C. roseus hairy root cultures
The plant expression vector p1304 + -Orca2, harboring the coding region of the wild-type C. roseus Orca2 gene and two selectable marker genes (hpt and Gus) on the same T-DNA fragment, was constructed (Figure 1) and transformed into disarmed Agrobacterium tumefaciens C58C1 strain carrying the plasmid pRiA4 of A. rhizogenes.Integration of Orca2 gene into the vector p1304 + and Liu et al. 3263 transformation of vector p1304 + -Orca2 into C58C1 strain was confirmed separately by PCR analysis and sequencing.Two weeks after infecting the leaf explants, hairy roots were induced from the wounded edges and surface of the leaves.Ten days later, the percentage of leaf explants to form hairy root lines was about 70%.
50 putative transgenic hairy root lines and 30 control hairy root lines were excised separately when they grew over 2 cm in length and transferred to the fresh regulatorfree half-strength MS solid medium that contained 500 mg/l cefotaxime to eliminate bacterial contamination.These hairy root lines were cultured in the dark at 25°C and were then routinely subcultured to the same MS solid medium every two weeks (Figure 2A).After two months of subculture, 38 independent putative transgenic hairy root cultures and 16 control hairy root cultures were obtained and transferred individually into the regulator-free half-strength B5 liquid medium for subculture.In comparison with growth on the solid medium, hairy roots cultured in the liquid medium grew more rapidly and had higher lateral branching (Figures 2B and C).However, the hairy root cultures changed gradually from white to red-brown during the subculture, and it could be observed that a few of red-brown substance was secreted from hairy roots into the culture medium after 3 to 4 weeks.As a control, adventitious roots excised from C. roseus sterile seedlings were cultured on the growth regulator-free half-strength MS solid medium, but these roots grew very slowly, did not branch and perished after 2 or 3-week subculture period.
Of the 38 putative transgenic hairy root cultures, only 12 cultures could be maintained in liquid media after one month of subculture.At the same time, 8 control hairy root cultures were obtained.These cefotaxime-resistant hairy root lines (12 putative transgenic and 5 control hairy root cultures) were evaluated for growth, integration and expression of Ri plasmid T-DNA genes, and alkaloid contents in dry hairy root samples.

Molecular analysis of the transgenic hairy root cultures
By using the genomic DNA from the putative transgenic and the control hairy root cultures as template, integration of the rol genes (rolB, rolC) and hpt gene into the genome of C. roseus hairy root cultures was confirmed by PCR analysis (Figure 3).As expected, it was demonstrated that three fragments, with lengths of 423, 622 and 812 bp corresponding to rolB, rolC and hpt gene, respectively, were amplified only from the putative transgenic hairy root cultures but not from the control hairy root samples.These results indicated that the rolB and rolC genes from the Ri plasmid of A. rhizogenes C58C1 and the hpt gene from the plant expression vector p1304 + -Orca2 were all integrated into the genome of transgenic C. roseus hairy root cultures.
Real time-PCR is the most sensitive method for quanti-  tation of gene expression levels.SYBR green I-based quantitative real-time PCR method was used to characterize the Orca2 gene relative expression status in transgenic C. roseus hairy root cultures.The results indicated that the expression levels of Orca2 gene in 12 putative transgenic hairy root samples were different.Among them, the samples No.4, No.7 and No.12 expres-sed the maximum level of Orca2 gene (3.646-,3.96-and 3.19-fold) in comparison with the control samples, respectively, while the sample No.10 expressed the lowest level of Orca2 gene in all hairy root samples (the value is about half as much expression as in the control samples) (Figure 4).The fact indicated that expression of Orca2 gene was inhibited in the transgenic hairy root sample No.10.After detecting the gene relative expression of Orca2 gene, the seven hairy root cultures with high expression level were used for HPLC analysis.

Relative mRNA Expression
In conclusion, molecular analysis experiments for the hairy root cultures proved that the T-DNA from the disarmed A. rhizogenes C58C1 strain containing vector p1304 + -Orca2 had been integrated into all the putative transgenic hairy root cultures.

GUS histochemical assay
To monitor transgenic expression in the hairy root cultures, histochemical GUS activity assays were performed with seven putative transgenic and five negative control hairy root cultures from cefotaxime-resistant hairy root lines.All the putative transgenic hairy root cultures were stained blue and different samples showed diverse GUS activity with different blue spots.The control hairy root samples showed no significant GUS activity (Figure 5).These results indicate that the Gus gene in the T-DNA had already been integrated into the genome of the transgenic C. roseus hairy root cultures.

HPLC analysis of TIAs accumulation in hairy root cultures
TIAs profiles (catharanthine, vindoline and VLB) of the putative seven hairy root cultures (samples No.1,2,4,6,7,8,12) with relative high level expression of Orca2 gene, and negative control hairy root cultures were determined by HPLC analysis in this study.The results showed that the putative transgenic hairy root extracts accumulated more catharanthine and vindoline in comparison Liu et al. 3265 with the control hair root cultures.However, VLB content could not be detected in the transgenic and control hairy root cultures by HPLC method.The average catharanthine content in the transgenic hairy root extracts was 4.7869 ± 0.59 mg/g DW; over 2.03-fold higher than that in the control cultures (2.357 ± 0.415 mg/g DW).The sample No.7 accumulated the maximum level of catharanthine content (5.986 ± 0.672 mg/g DW) among the seven samples, while the sample No.4 accumulated the lowest level of catharanthine content (3.815 ± 0.376 mg/g DW) (Figure 6, upper panel).The average vindoline content in the putative transgenic hairy root cultures were 0.144 ± 0.0157 mg/g DW, over 3.67-fold higher than that in the control cultures (0.0392 ± 0.0054 mg/g DW).It is noteworthy that the transgenic sample No.2, but not the sample No.7, accumulated the highest level of vindoline content (0.1888 ± 0.0024 mg/g DW) (Figure 6, lower panel).

DISCUSSION
Transcription factors are regulatory proteins that modulate the expression of specific groups of genes through sequence-specific DNA binding and protein-protein interactions.They can act as activators or repressors of gene expression, mediating either an increase or a decrease in the accumulation of messenger RNA (Broun, 2004).More than ten transcription factor genes have been cloned and characterized in C. roseus till now.Among them, Orca1, Orca2 and Orca3 were characterized to control closely the expression of some genes involved in TIAs biosynthesis.ORCA1 had been approved to be expressed constitutively and had little effect on Str promoter activity.ORCA2 activated the Str promoter and its expression was rapidly inducible with jasmonate (JA) and elicitor (Menke et al., 1999) are enhanced by overexpressing ORCA2 in C. roseus hairs roots.It had been approved that catharanthine is distributed equally throughout the aboveground and underground tissues of C. roseus.However, vindoline, tabersonine as well as the dimeric alkaloids are restricted to leaves and stems because of the late steps of vindoline biosynthesis which require specialized cell types, idioblast and laticifer cells, which are located in stems and leaves (Westekemper et al., 1980;Deus-Neumann et al., 1987).Therefore, many cell and hairy root cultures produce catharanthine and tabersonine, but do not produce vindoline because there is a limitation in the conversion from tabersonine to vindoline (Shanks et al., 1998).This fact had been confirmed by some experiments (Bhadra and Shanks, 1997;Brillanceau et al., 1989;Toivonen et al., 1989).
In this study, however, the results demonstrated that the transgenic hairy root extracts accumulated 2.03 and 3.67-fold of catharanthine and vindoline more than the control hairy root cultures, respectively.Previously, Parr et al. (1988) reported a similar result: an experimental detection by immunoassay demonstrated that ajmalicine, serpentine, vindolinine and catharanthine were prominent components in C. roseus hairy root cultures during all stages of the growth cycle.Vinblastine could also be detected by a combination of HPLC and radioimmunoassay, though at a low level (0.05 g/g DW).They speculated that the differentiated characteristics of the hairy roots have offered the potential for the production of various monomeric indole alkaloids.It can be proposed that the main reason for these results is the activation of the enzymes in vindoline biosynthesis pathway in hairy root cultures by weak light during their subculture in liquid medium, since it was found that the hairy root samples had been lightened by some weak scattered light around in the culture room.It was noticed that some early experiments had already proven that light plays a critical role in TIAs biosynthesis in C. roseus plants during their growth and development.Further studies showed that phytochrome is involved in the activation of the last two enzymes in vindoline biosynthesis; D4H (Vazquez-Flota and De Luca, 1998) and DAT (Aerts and De Luca, 1992), in Catharanthus seedlings.In addition, Ramani and Jayabaskaran (2008) reported that catharanthine and vindoline increased 3 and 12-fold, respectively, on treatment with a 5-min UV-B irradiation in the suspension cultures of C. roseus.The other possible reason for the results is that the overexpression of ORCA2 in the hairy root might have triggered the flow of TIAs pools towards AVLB biosynthesis in which catharanthine and vindoline are needed as precursors.Therefore, our next research would focus on investigating the exact mechanism of light on the vindoline biosynthesis in C. roseus hairy root cultures and on the metabolic flow during TIAs biosynthesis in C. roseus.

Figure 2 .
Figure 2. The induction and subculture of transgenic hairy root cultures; (A), Hairy roots were induced from the leaf explants C. roseus and were cultured on MS solid medium; (B), hairy roots growing in half strength B5 liquid medium in a 250 ml Erlenmeyer flask for ten days; (C), hairy roots were cultivated in the liquid medium for one month.

Figure 3 .
Figure 3. PCR analysis for the presence of rolB, rolC and hpt gene in independently transgenic hairy root cultures.M, DL2000 marker; lane +, plasmid p1304 + -Orca2 was used as positive control; lane -, untransformed C. roseus root DNA was used as a negative control; lanes 1 to 12, 12 individual hairy root cultures transformed with Orca2 gene.

Figure 4 .
Figure 4. Relative mRNA expression levels of Orca2 gene in transgenic hairy root cultures of C. roseus checked by real-time PCR method.The mRNA expression value in untransformed control sample was 1.0; O2-1 to O2-12, 12 individual hairy root cultures transformed with Orca2 gene; Data shown are means ± standard deviation of three replicate measurements.